Transcription
I used to say that NotebookLM from Google and projects in ChatGPT were pretty much the same. You're dump in a bunch of stuff that you're working on or learning, and then you can ask follow-up questions, make quizzes, whatever. I don't think I can say that any more.
Why? Because Google has really focused on the learning experience in NotebookLM, making it a mostly dedicated learning portal for whatever you're learning. They've added a bunch of new features since I tested it a little less than a year ago. And I want to see if those features are good. If they actually make the learning experience better.
So let's see what's new, what's fresh in NotebookLM. I'm going to be using the same notebook that I created last time I reviewed NotebookLM. I guess that was 218 days ago. This notebook is a sentence patterns course, basically a grammar course. It's one of my courses, and I think it's a good one to use because I know the content very well. So I know what's accurate and what's not.
Now, what looks new here? Well, a lot of things. The interface looks a bit different, but there are more options here, more things you can do with the content. So if you haven't seen that review, basically what I did was I uploaded all of the PDFs associated with the course that come along with the course. Different aspects of English grammar, gerunds and infinitives, hypotheticals, open questions, sentence beginnings, active and passive, a bunch of things. I also added a YouTube video, one of my YouTube videos about gerunds and infinitives, sort of to round out that specific topic. And I think there was a summary before, but maybe not. You know, it's been a while. I have to go back and review that review. I don't remember if there was a summary or not, but this is a decent summary and there are some key topics. Gerunds explained, infinitives explained, usage examples. So this is just for exploring each source, right? Clicking on each specific source, PDF, video, whatever. You get key topics, you get a summary, right, and you get the content of the source. And is there a full transcript in this? Is there a full transcript? Yeah, I think that's the full transcript below, the entire transcript of the video, which is maybe a benefit of using a tool by Google because YouTube is also by Google. If you're uploading YouTube videos or adding YouTube links, it can easily just pull the entire transcript. That's cool.
So then this is the chat interface. This is pretty much the same. But I want to try a few of these things. Before, I think we tested out the podcast feature. We made some quizzes. I wasn't too impressed with the quizzes. They were fairly accurate, but a little bit clunky. And I remember being more impressed with ChatGPT's quizzes in projects. But the podcast feature is kind of the famous feature of NotebookLM, right? And I think that's cool. But after you do five of them, I don't know, maybe it's just me. They start to feel a little repetitive. And so one thing I really wish they would do is add more voices and more characters. I don't want to hear the guy and the lady basically talking like this. "Hey, oh, that's so cool. This is interesting." Every voice that you hear from Google or OpenAI, they sound like a millennial, a spoiled millennial from California, with vocal fry and Valley Girl English. Anyway, that's just a little pet peeve.
So we're going to try that out, but there's something different about it that I really want to focus on, which is that it's interactive. So you can kind of jump in and say, "Hey, can I ask a question?" Or, "Hey, I'd like to add a point here." So that's cool. And we're also going to try the mind map, which wasn't there when I tested it before. And the video overview, which is pretty cool. And if I can move aside here, study guide, FAQ, I think, was there when I tested it. And the timeline, I'm not sure that's very relevant.
So let's start with the mind map. You just click on mind map and it goes. Mind map go. Can I just do all of these at the same time? Let's see. Let's customize the audio overview. I want to see how much I can influence this, especially the style, the voices. So let's try. Let's try this. I want to hear two people with British accents, let's say, or non- or just not American accents, discussing each each main grammar point briefly with one person in a tutor role and one as the student. Okay, so let's see if we can get that. Again, I want to see how interesting we can make it and how much variability we can get.
All right, the mind map is ready. So here we go. This looks very cool. Okay, we can zoom in and out. English grammar concepts. We have, okay, active and passive voice, questions, commands, requests, conditional sentences. These headings all look correct. Again, the reason I chose this one is because this is a course I created. This is one of my courses. By the way, this course is part of the 90-day program. This is one of the five courses in the 90-day program. It's more than just grammar. It's learning how to structure sentences, think in English, not just writing, but how you make different kinds of sentences when you communicate in English. I'm very proud of this course. I think it's an excellent course, and I think it teaches you how to program your brain. It's one of the courses that unlocks as you go. Anyway, enough about that.
So, let's see. Exploring voices. Okay, and then we explore voices with specific examples. Okay, it doesn't give me the breakdown of which one is active and which one is passive, but it does give the examples provided. I think those are pretty much right. Let's go to conjunctions. Yes, broken down into coordinating and subordinating. And then you have all of the coordinating and all of the subordinating, or at least the ones I think that we focused on. That's pretty cool. Transformations. Let's do hypotheticals. Let's look at hypotheticals. So this is an interesting way to just kind of make sure you've got everything, right? A good way to go back and say, "Hey, am I missing anything? Hey, is there anything I can quickly look at again and just get a refresher, right? A quick refresher." So having this structure, I think, is very helpful.
So once you get to this, it doesn't break down further. But okay, when you click on it, then you get an automatic prompt in the chat. "The sentence you may be struggling from PTSD as so many others do if you hadn't been emotionally ready for what you saw is presented within the past section of the hypothetical's source. In the context of the past, this sentence illustrates a mixed..." and then it goes into "Mixed conditionals, past conditional, present result." Okay, that's a really interesting breakdown and adds very useful context, a kind of extension on top of what you're already learning. So I would encourage, you know, anyone currently in, let's say, the 90-day program, as you go, as you get PDFs along the way, of course, it's the videos too, put them into NotebookLM, get these deeper breakdowns, explore, play, make more quizzes. This encourages you to do. And that's the key thing, I think, with this, is this is encouraging you to interact with what you're learning even more, right? With a 90-day program, you have opportunities to practice what you're learning with assignments, right? And you're also getting feedback from AI and having conversations. Plus, you have my WhatsApp. But also, adding this to it would be a fantastic way to explore further, to extend, to follow your curiosity. And I can download it. That's pretty cool.
Let's look at, while the audio overview is going, the video overview, something that's, I think, also relatively new. Let's see. You know what? I don't want to constrain it because I don't know. I don't know how much I can control the output. I want to see what it creates. So I'm just going to click on it and see what it can do. This is the previous FAQ that we created. I think they had this feature before. The thing I remember about this is I wasn't crazy about the formatting, right? This sentence is just visually broken in half, looks a bit awkward. The formatting is just, you know, it looks looks like a wall of words. It's just a bit awkward to go through. Although the content is decent. So I'm curious if it will look the same if we do another one. So let's do an FAQ and let's do a study guide and just see what happens.
In the meantime, I'm going to just do a few queries here. "Can you explain the different types of conditionals?" See what we get here. "The sources outline several types of conditional sentences, each with a specific structure and use: Zero, First, Second." Very good citation, always citing exactly where it comes from. But the nice thing is I don't have to actually look at the original, for example, PDF. It gives me the accurate text, right, in a in a format that's easy to look at, just hovering over the number for the citation, right? "Implied conditional structure: The if clause is implied rather than explicitly stated. Uses: The conditional is assumed. Example: Had I known I would have helped (implied: if I had known, if I had known)." Yes, yes. Okay, that's great. That is fantastic, and I can save that. Okay, so I'll save to note. Save that. This is kind of your timeline of events, basically.
I see the video overview is done. Now, one thing I want to do quickly before we look at that is just a quiz. "Create a quiz a very difficult quiz about active and passive voice." Now let's take a look at some of the things we've got here. The video overview. Oh, the video overview is still going. It's the audio overview that's done. And one thing I want to see is how interactive it actually is. So if I, I don't know why I'm here. Interactive mode allows you to say stuff as they go. So let's try it out. I'm going to move myself. That's better. That's way better. Let's try this out.
"All right, have you ever found yourself pausing mid-sentence, perhaps wondering if you've got your witches and whats in quite the right place? Or, you know, if you're truly conveying what you mean with the precision you intend? It's something lots of us run into, isn't it? Today our deep dive is all about English grammar, not as some rigid, dusty set of rules from school, but actually as a surprisingly flexible and incredibly powerful toolkit for clear, impactful, and often quite, well, beautiful communication. We're going to be drawing from a collection of concise, really practical guides on some fundamental grammar points that you, our listener, have shared with us. Our mission, as ever, is to extract the most important nuggets of knowledge from these sources, turning potential confusion into those lovely 'aha!' moments. We're going to tackle this with a slightly different format today, a sort of Q&A tutor-student style, if you will, to really hone in on what you might need most. Our goal is to give you a genuine shortcut to being well-informed."
So a couple things. One, I think I noticed in every single podcast I generated, or audio overview, I guess we're calling it, they said "aha moment." It's almost like it's baked into the system prompt: "Make sure to use 'aha moments' somewhere." They used it again. Also, the voice is definitely not a British voice. That's an American accent. That's a guy from California talking about grammar. And maybe okay, but at this point, I would really like a lot of customization. I would like to be able to control. And you can do this in ElevenLabs if you use ElevenLabs. But here, as a learning tool, I want to be able to control the style of conversation. I want to control the voices, the intonation. I want to control maybe how many people are there. I really want control, control over this. And if they did this, if they made that way more robust, that would be extremely powerful. That's my main criticism at this point. But, you know, still useful about core grammar principles, packed with surprising insights, and, naturally, whoa, did someone just break the fourth wall? Let's hear it.
"Yeah, I think I just broke the fourth wall."
"Can you hear me okay?"
"You absolutely did. And we hear you loud and clear. Welcome. Yes, that's fantastic. We love it when you chime in. You know, it's actually a great example of what we're talking about today."
"Indeed. What's fascinating here is how often seemingly so..."
"Hey, yeah, what's up?"
"Well, yes. So first, I'm wondering if you could get to the point. And second, not to be rude, and second, maybe maybe you could really focus on the examples because I'm not so interested in the theoretical, you know, names of things and all that."
"That's a great request. And we absolutely can. We want this to be as helpful for you as possible. No rudeness at all. We love the feedback. Seriously. You know, when we talk about this being a Q&A style show..."
"Oh, go for it. Sorry, before we continue, can you guys both do like a goofy noise? I want to hear each of you do a silly noise like boop, boop, boop, something like that."
"Ah, a request for some levity. I appreciate that. You know, it's interesting because sounds and interjections are a form of communication, too. And while a goofy noise isn't strictly grammatical, it's certainly conveys theoretical discussion about..."
"Doing. Just do the goofy noise."
"Exactly. It's a command in a way, or a request. Like, 'go away' or 'get over here,' but, you know, much sillier. So do it."
"My goofy noise, if I had to pick one, would be a very thoughtful, hmm."
"Okay, my turn. Mine would be a sort of enthusiastic boing."
"So we covered the goofy noises. Q&A style show. This is exactly what we hope for. You're right, the theoretical can sometimes obscure the practical application. So we were just saying how the passive voice shifts emphasis. Let's jump straight to a concrete example that highlights this. Take the classic, 'My leftover pumpkin pie was taken by someone.' We mentioned it before, but let's really look at it. Here, the pie is the subject, receiving the action. The focus is squarely on the pie."
"Yes, absolutely. We can both jump in with examples."
"All right. So I feel like let's prioritize getting to the point. They're talking a lot about the thing I asked them to do. And it seems like they're talking about the thing I asked them to do rather than doing the thing I asked them to do. And they've got the same sort of things in there, the same voices, the same intonation, the same style. I'm not so crazy about any improvements they may or may not have made to the podcast or audio overview. It's, again, it's fine. It's useful. I just wish it were way more customizable."
Now let's check out, this is what I really want to see. Now let's check out the video overview. Oh, yeah, we should see how good the quiz is. Part two, transformations. "Transform the following active voice sentence into passive voice, ensuring the agent is explicitly mentioned as per the sources: Someone took my leftover pumpkin pie." My leftover pie was taken. My leftover pie was taken by someone. My leftover pie was taken by some nefarious unknown food thief. Someone's pie was taken by me. I'm not sure if that's a great question, actually. Hmm. No, that isn't a great question. If you were going to do this question, you'd have to make the answer, the possible answer is a bit different. Okay, it's giving the source, but no, this is not a great quiz. And same issue as last time. It has highlighted the answer. So a quiz is supposed to be a quiz, not some possible choices with the correct answer highlighted. So not great there. That's disappointing. That's very disappointing.
So they've added these features, but have they actually made the deeper intelligence? Because that's kind of what I didn't like as much last time, the quality of the quiz is. The accuracy was pretty good, but the quality of the quizzes, the overall quality of the responses. The formatting was my main complaint when I first reviewed NotebookLM, and eh, I'm not blown away. Let me say that.
So let's check out the... But I am blown away by the mind map. I think the mind map is incredibly cool. And legitimately very useful. The FAQ looks similar, similar formatting, yes, the large space is there. I just wanted to see if the formatting was better, and the formatting is not better. And this is the study guide. "Sources outlined several types of conditional sentences." Okay, not just conditionals. Why is this just giving conditionals? Hmm. Okay, I'm not sure that's super useful. That's just one of the topics, one of many. But okay.
What I really want to see is this video overview. The hot new feature. Where is it? Is this it? Is this the video overview? I think, oh, this is the video overview. Okay, here we go. Six minutes and 17 seconds. We won't watch the whole thing, but let's just see if it's any good.
"Hey, welcome to the Explainer. Today, we're doing something a little different. We're not just going to look at sentences. We're going to become sentence architects. Yeah, that's right. We are going to uncover the actual blueprints for building powerful, complex, and genuinely interesting sentences right from the ground up. So let's just dive right in. Take a look at this. On the left, you've got a simple fact: Jesse laughed. It's clean, it's simple, it gets the job done. But on the right, oh man. On the right, we have a full-blown comedy scene. There's a hat, a terrible haircut, squirting milk, and uncontrollable giggling. So how do you get from that simple statement to this incredibly detailed story? That journey, well, that's what being a sentence architect is all about. Let's take a look at the blueprints. Okay, so every great structure has to start with a solid foundation. For us, as sentence architects, that very first choice is absolutely fundamental: who or what is the star of our sentence? Who's in charge of the action? This brings us to our very first tool: the choice between active and passive voice. This whole decision, it really just boils down to one simple question: Where do you want to shine the spotlight? In the active voice, the subject is the hero of the story. 'I cleaned the kitchen.' I did it. But in the passive voice, the focus totally shifts to the object that received the action. Suddenly, the kitchen is the star of the show. Check this out. Here's the active voice in action. It's direct. It's crystal clear, and there is zero question about who the culprits are. The sentence puts the focus squarely on the teenagers and their act of, well, vandalism or public art, I guess it depends on your perspective. Now watch what happens when we switch to the passive voice. Poof! The teenagers have just vanished. All of a sudden, the bell itself is the subject. The focus is on the result of the action, not the car. And this is a really powerful architectural choice, especially if you either don't know who did something, or maybe you don't want to say. Of course, you can still bring the doer of the action back into a passive sentence. You just use the word 'by.' But notice the difference in emphasis, right? The teenagers are now tacked on at the end, almost like an afterthought. The bell, it remains the primary focus. So, we've laid our foundation, but a foundation by itself doesn't make a house, right? Now we've got to connect our ideas. This is where conjunctions come in. For a sentence architect, these are like the mortar, joining individual clauses and ideas into stronger, way more complex structures. And these guys, these are some of your most essential tools. Each one of them creates a totally different relationship between ideas. You know, and just adds on more info. But, and yet, create some friction, some contrast, and so builds this really nice logical bridge from one point to the next. Here's a super simple, classic use of but. It takes two separate ideas: 'Thai food is great,' and 'Japanese food is better,' and just locks them together in a direct comparison. Instantly, the sentence doesn't just state facts. It has a point of view. And look at how so works here. It establishes a really clear cause and effect relationship. Because of the first part of the sentence, the second part happens. It's that logical connection that really makes an argument flow. All right, the structure is up, and the pieces are all connected. Now comes the really fun part: adding the details that make your sentence totally unique and memorable. This is where we, as architects, get to start painting with words. You know, one of the easiest ways to add some real style to your work is to just change how you start your sentences. Instead of always starting with the subject, you know, 'I did this,' 'the bear did that,' try leading with a description. It creates this nice rhythm and a little suspense, and it immediately makes your writing way more dynamic."
Okay, I have to say, that is pretty impressive. That is very cool. I mean, first of all, visually, very clean. When I've asked ChatGPT agent to create a slideshow, I've tested it out quite a bit, a slide deck, right? PowerPoint. It looks like crap, right? It's good. It's useful. I mean, all the information is there, but it doesn't look good. This looks good. This has emojis and it's got a color sort of theme. And I wonder how customizable the color scheme is. I'm going to have to play with it more. But the visuals make sense. The flags, when talking about Thai food and Japanese food, were the Thai flag and Japanese flag. So visually, it's kind of reinforcing the content.
Now, I will say to condense so much into a short video is skipping so much that it would be really hard to really get each one to really get the point of each one. In order to really get it, you need more explanation. You need more examples, right? You need to look at it from different angles. So while it's very well explained, I think for the next one, what I would do is ask for it to focus on one particular topic. Focus on conjunctions. Focus on active and passive voice. Focus on gerunds. Focus on hypotheticals. Focus on sentence beginnings. Just do that. And I might create a standalone video for each focus. I want to see how far I can drill in and how deep it can go, how in-depth it can explain. Because this just flying through, it gives you a general sense, right? It gives you a feeling for how an architect of sentences put sentences together. But it would be hard to, I think, deeply understand all of the pieces, right, flying through at this speed with an overview this broad. That said, I think for what it is, it's really good.
Now, again, like with the audio overview, what I want to make it not a gimmick. And to be clear, I actually don't think this is a gimmick. I think this is better than the audio overview. The audio overview is, in my opinion, okay, it's too roundabout. It's too not really talking about the pith and the depth of it and too much jokey conversational vibes. This is talking about the actual content, stitching it together in a way that seems to flow and makes sense. So I want maybe more of this substance in the audio overview. I think this is a bit better, but I want more control over both when I prompt. Again, I want to control the voice. I want to control the style. I want to control the depth and the detail. I want to control how much freedom I give or if I want it to be very strict about sticking to the sources, right? I want to have full control over that. I need to, of course, do a bit more testing, but the video overview, that is very, very cool. That is actually amazing. I just wish I could control the voice. I wish I could change the style a bit. She uses very kind of modern speech, and I would want her to simplify a bit, use simpler language, but it's hard to complain. I think that level of customization, plus working on the deeper intelligence of, let's say, the model that's behind the chat, would take this into a level of usefulness that I haven't seen before, that would be honestly incredible.
So, you know, no quizzes like this. Real deep intelligence behind the chat. I'm still not really seeing that same level of intuition that I get maybe with ChatGPT, right? When I use ChatGPT to talk about projects that I'm working on. And then a lot of customization. If they improve those things, this is this is an amazing platform for learning. Because again, it makes you interact with what you're learning. It's important to do stuff when you're learning. Some people get this idea: "If I just sit back and listen to a course, I'll be better." And that's just not the case. You have to actually do stuff. You have to put in the work. You have to try. You have to make examples. You have to get feedback. You have to interact. Input plus output. That's the key, right? And if you, if you don't do that, then it's not going to stick. You're not going to really, really get it, right? So I think that's the best thing about NotebookLM, that it encourages you to play and interact. I just wish it would continue to get smarter. I wish all of the features were deep and nuanced and powerful and customizable. That's it. Let me know what you think. Thank you for watching. If you haven't done so already, make sure to hit the like button and subscribe. Get a free course in the links in the description, check out the seven-day challenge, and the 90-day program, which includes this course as one of five courses, over 50 hours of content. It's also going to push you to do stuff. You have to do assignments, practice with AI. There's a lot of interaction. You have my WhatsApp throughout the program to help you along. You don't have to complete it in 90 days. You can take longer if you need to, but there are a couple of things that you get if you do complete it in 90 days. So make sure to check that out. Links in the description. Thank you for watching again, and I will see you in the next one.